ABSTRACT As one of the major potential alternatives to animal models, the read-across of toxicity data within groups of similar compounds represents a promising direction to fill the data gap in chemical safety assessment. While read-across can play a key role in complying with legislation (e.g. the European REACH regulation), most of the current read-across tools only rely on chemical structure information. With more and more available biological data, read-across based on big data can add extra strength to this process. In this project, we will develop an automated computational approach 1) to explore the public big data sources and generate the bioprofiles; 2) to perform a read- cross study using the target 10,000 compounds with animal acute toxicity data; and 3) to reveal the potential toxicity mechanisms from the public biological data. Then the external acute toxicity database, which contains around 5,000 new compounds, will be used to validate the resulting read-across approach. Moreover, we will share with the toxicology community the developed read-across tool via Chemical In vitro-In vivo Profiling (CIIPro) portal (ciipro.rutgers.edu), which has been proven to be a useful toxicity evaluation tool by US EPA. The toxicologists can use the CIIPro portal to directly evaluate acute toxicity of new compounds avoiding animal testing; to illustrate the relevant toxicity mechanisms; and to prioritize toxic compounds of environmental interest for future animal studies.